Dmitriy Gizlyk / 프로필

In previous articles, we used stochastic gradient descent to train a neural network using the same learning rate for all neurons within the network. In this article, I propose to look towards adaptive learning methods which enable changing of the learning rate for each neuron. We will also consider the pros and cons of this approach.

We have previously considered various types of neural networks along with their implementations. In all cases, the neural networks were trained using the gradient decent method, for which we need to choose a learning rate. In this article, I want to show the importance of a correctly selected rate and its impact on the neural network training, using examples.

We have earlier discussed some types of neural network implementations. In the considered networks, the same operations are repeated for each neuron. A logical further step is to utilize multithreaded computing capabilities provided by modern technology in an effort to speed up the neural network learning process. One of the possible implementations is described in this article.

We continue studying the world of neural networks. In this article, we will consider another type of neural networks, recurrent networks. This type is proposed for use with time series, which are represented in the MetaTrader 5 trading platform by price charts.

As a continuation of the neural network topic, I propose considering convolutional neural networks. This type of neural network are usually applied to analyzing visual imagery. In this article, we will consider the application of these networks in the financial markets.

In this second article, we will continue to study neural networks and will consider an example of using our created CNet class in Expert Advisors. We will work with two neural network models, which show similar results both in terms of training time and prediction accuracy.

Artificial intelligence is often associated with something fantastically complex and incomprehensible. At the same time, artificial intelligence is increasingly mentioned in everyday life. News about achievements related to the use of neural networks often appear in different media. The purpose of this article is to show that anyone can easily create a neural network and use the AI achievements in trading.

This article is a follow-up to the previous one called "Reversal patterns: Testing the Double top/bottom pattern". Now we will have a look at another well-known reversal pattern called Head and Shoulders, compare the trading efficiency of the two patterns and make an attempt to combine them into a single trading system.

Traders often look for trend reversal points since the price has the greatest potential for movement at the very beginning of a newly formed trend. Consequently, various reversal patterns are considered in the technical analysis. The Double top/bottom is one of the most well-known and frequently used ones. The article proposes the method of the pattern programmatic detection. It also tests the pattern's profitability on history data.

Using limit orders instead of conventional take profits has long been a topic of discussions on the forum. What is the advantage of this approach and how can it be implemented in your trading? In this article, I want to offer you my vision of this topic.

The main advantage of trading robots lies in the ability to work 24 hours a day on a remote VPS server. But sometimes it is necessary to intervene in their work, while there may be no direct access to the server. Is it possible to manage EAs remotely? The article proposes one of the options for controlling EAs via external commands.

Efficiency of any trading robot depends on the correct selection of its parameters (optimization). However, parameters that are considered optimal for one time interval may not retain their effectiveness in another period of trading history. Besides, EAs showing profit during tests turn out to be loss-making in real time. The issue of continuous optimization comes to the fore here. When facing plenty of routine work, humans always look for ways to automate it. In this article, I propose a non-standard approach to solving this issue.

There are numerous trading strategies out there. Some of them look for a trend, while others define ranges of price fluctuations to trade within them. Is it possible to combine these two approaches to increase profitability?

Comparing several time series during a technical analysis is a quite common task that requires appropriate tools. In this article, I suggest developing a tool for graphical analysis and detecting correlations between two or more time series.

The trade Signals service develops in leaps and bounds. Trusting our funds to a signal provider, we would like to minimize the risk of losing our deposit. So how to puzzle out in this forest of trade signals? How to find the one that would produce profits? This paper proposes to create a tool for visually analyzing the history of trades on trade signals in a symbol chart.